Table constructor
Key features:
Main part of prompting implemented at the code level
Zero 'hallucination' in AI responses when creating Markdown syntax
Compare different AI models' responses on one page
Multiple AI models' responses from a single request (if one model gets stuck, Explorers always check results from other models)
Unlimited quantity of relevant to each other separate tables in a single AI response
All fields of the Table Constructor form are connected to AI
Supports multiple data types
The Table Constructor function offered by Dataexplorers.ai provides a straightforward and efficient approach to handling structured data. Additionally, it supports multi-user mode, enabling a flexible way for read and write operations.
The Table Constructor is a component responsible for providing data to an AI in JavaScript Object Notation (JSON) format. The AI system processes the input data and generates a response, which is then rendered in a tabular format using Markdown syntax.
Essentially, the Table Constructor is like a prompt implemented at the code level to help users focus on creating simple and concise prompts without multiple descriptions of table formatting. Additionally, it allows for experiencing zero 'hallucination' in AI responses provided in Markdown syntax.
Markdown is a lightweight markup language that allows formatting plain text documents with headings, lists, links, and other formatting elements. The tabular format in Markdown provides a clear and concise way to present structured data, making it easy to read and interpret.
Here's an example of how structured data, like news event type information, could be represented in a Markdown table:
This approach leverages the strengths of JSON for data interchange, AI for data processing, and Markdown for structured data representation, providing a powerful and flexible solution for managing and displaying complex data sets.
Dataexplorers.ai table constructor function provides you with the following query modes:
Table query mode Explorers can read/write/modify/delete any data in the table.
Fields: Each field contains the following configurations:
1. Field: The name of the field, which must meet the following requirements:
Can only contain letters, numbers, or “_”.
2. Prompt: A prompt for the field.
3. Data type: The data type of the field. LLM will process and save the content submitted by the user according to the data type. Currently, the following data types are supported: String, Integer, Time, Number, and Boolean.
4. Mandatory: Indicates whether the field is necessary. A necessary field means that the row must appear in the AI response.
5. Sources feed relation: If this switch is ON, the AI will process data exclusively from the sources.
6. Actions: To remove a template, please use this token: ae4u7uk96hzVWRKujCjauOJZkptTnsWRZKzgC3
Use the table: dataexplorers.ai supports managing data in the table in prompt and Explorers queries, or through SQL statements of a database in a workflow.
To facilitate the demonstration of the dataexplorers.ai database, a table with the following data is used as an example in this document:
Note:
Please fill in each configuration item carefully. Large language models will use these configurations to match user inputs. Proper naming can improve the semantic matching of LLMs.
Delete and Edit a Table:
Explorers can delete or edit an existing table.
If a field name is changed, existing data will be stored under the new field name.
If a field is deleted, the data associated with that field will also be deleted.